Best under-desk cardio with adjustable resistance

“Adjustable resistance” is the second thing every listing in this category says about itself, after how quiet it is. It is also, on most of these machines, the least substantiated claim on the page.

We track 6 seated under-desk machines and 3 standing mini steppers. 5 of those 9 publish a resistance level count on the maker’s own site or the retail listing we link. 4publish no count at all — not a vague one, not an unsourced one, none.

Then there is the number underneath the number. Exactly one machine in this category publishes what its resistance is worth: the DeskCycle, whose maker states eight magnetic settings up to 39 lb of resistance. Every other count on this page is a bare integer. You are told a range has been divided into eight, or six, or sixteen parts, and never what the parts are parts of.

That is a strange gap, and it is worth sitting with rather than working around. A resistance level is the only control on most of these machines. It is the entire adjustable surface of the product. And for four makers out of five who print a count, the unit is missing.

This is the part that changes a buying decision, so here it is plainly. A level count tells you how finely a range has been divided. It does not tell you how wide the range is, and it does not tell you where the top of it sits.

Put the two published counts side by side. The YOSUDA publishes 16 magnetic levels. The DeskCycle publishes eight. The obvious reading is that the YOSUDA is twice the machine, and its own listing invites that reading. But the DeskCycle is the one that tells you its eighth setting is 39 lb of resistance, and the YOSUDA does not tell you what its sixteenth is. Sixteen steps could divide a narrower total range more finely than eight steps divide a wider one. From the outside there is no way to know, and nothing in either listing closes the gap.

So the useful way to read a level count is as adjustment resolution, which is a real benefit and a smaller one than it looks. If you are pedaling while you read, the thing you want is to nudge the effort without thinking about it, and more steps make that nudge smaller. What more steps do not do is raise the ceiling.

We rank the published magnitude first for that reason. It is the only figure in this category that survives being compared between two brands.

The second thing worth having is the mechanism, because it tells you what the resistance will do over time rather than how much of it there is.

• Magnetic is non-contact. A magnetic field slows the flywheel without touching it, so there is no friction surface to wear down and nothing that starts to squeal at the top of the range. Four of the nine machines here state magnetic resistance outright: the DeskCycle, the YOSUDA, the Cubii JR1 and the Cubii Move.
• Hydraulic cylinders are how the steppers work. Niceday publishes that its stepper uses hydraulic cylinders, and the Xiser publishes three cylinder settings adjusted by a knob. This is a different kind of resistance: you are loading a cylinder with your body weight from above, so the setting changes how quickly it yields rather than how hard a brake grips. It does not belong on the same scale as a pedal machine’s levels, and we do not put it there.
• And then there is “adjustable tension”. The Stamina InMotion E1000 is, on almost every other measure, the best-documented machine in this category — it is the only seated machine whose maker publishes dimensions, unit weight and a maximum user weight together. On this one specification it says only that the tension is adjustable. No count, and no statement of whether that tension is magnetic or contact. On a page about resistance, that omission is the finding.

A mechanism you have been told is worth more than a count you have not. That is the order the ranking uses, and the Stamina is the machine it costs the most.

One practical warning before the picks, because it is the way this particular search goes wrong.

The machine with the most resistance levels on this page, the YOSUDA, publishes no pedal height at all. Pedal height is the figure that decides whether your knees clear the underside of your desk through the whole stroke, and it is the first number we sort on when we rank these machines on fit. A machine you cannot get under the desk has no usable resistance range whatsoever, however finely it is divided.

So if you are shopping this cluster, measure first and shop second. Our desk clearance guide sets out the two measurements that matter and what the published heights in this category actually clear. The DeskCycle and the Cubii JR1 publish 10 inches; the Cubii Move publishes 9.5, the lowest figure we have found in the category; the Stamina publishes 11.38, the tallest. Two of the nine publish nothing.

A good share of the searches that land on a page like this one are actually looking for elastic bands rather than a resistance mechanism, so it is worth separating cleanly.

Bundled elastic bands are an upper-body accessory, not the machine’s resistance. They do not have levels. Their resistance is whatever the band you clip on provides, they are not connected to the flywheel or the cylinder, and they work your arms while your legs do something unrelated.

We do not credit a machine on this page for shipping with them, and the reason is methodological rather than snobbery: listings in this category are inconsistent about whether bands are in the box, so ranking on bands would mean ranking on a bundle that can change without the product changing. If bands are what you want, they are an accessory purchase that works with any machine here, including the ones that come with none.

Read the resistance column against the capacity column rather than on its own. The machine with the most levels publishes no maximum user weight and no pedal height, and the machine with the fewest publishes the lowest height in the category — which is the trade this whole page is about.

Ordered by the criterion at the top: published magnitude first, then published mechanism, then the level count.

We have not measured the resistance of any machine on this page. We do not own them, we have no force gauge and no instrumented flywheel, and we have not turned a single one of these knobs. Every figure above is read from a manufacturer or retailer page named in the sources below, with the date we read it. The reasoning about what a level count can and cannot tell you follows from what a count is, not from testing.

We have not converted anybody’s level count into a magnitude. It would be easy to assume that sixteen levels spans twice the range of eight, or that every maker’s level five is roughly comparable, and both would be inventions. Where a magnitude is not published we say it is not published and leave the cell empty. Our full method.

We have not ranked the steppers against the pedal machines on resistance. A hydraulic cylinder loaded by body weight and a magnetic brake on a flywheel are different mechanisms measured in different terms. The Xiser appears on this list because it publishes a count, and it is placed below the pedal machines rather than interleaved with them on purpose.